The control performance of an air-conditioning system is assessed using a qualitative method of evaluation. Fuzzy logic is used to relate performance criteria expressed in the form of IF-THEN rules to quantitative measures of energy consumption, discomfort, and maintenance costs. Test data were generated using an emulator consisting of a real-time simulation of the building shell and HVAC plant, together with a hardware interface that connects the simulation to commercial control equipment. Two case studies are presented. In the first, the effect of changing the strategy used to determine the zone temperature set-points is evaluated using 'expert rules', generated by a hypothetical facilities manager. In the second case study, the effect of varying the tuning parameters of the control system is evaluated using two sets of rules assumed to represent the differing perspectives of a facilities manager and a control engineer.